At SAP, AI investments are starting to pay off. Users of the S/4HANA ERP can use AI tools such as Joule to work more efficiently, for example, for transaction analyses and automated actions. The most significant gains are currently visible in service management, financial processes, and logistics chains. SAP is also working on a framework where multiple AI agents can work together. AI will soon be doing work in just about every area. We spoke with Eric van Rossum of SAP about the state and future of AI in cloud ERP.
Around 80 percent of business processes could already be supported by an advanced form of AI. End users can navigate through systems with the help of assistant Joule, that can understand the business by accessing data from SAP systems. Its versatility is its strength, from calling on Joule’s help to write an ERP extension to drafting a new job posting from the HR department.
At the same time, the entire software industry is moving toward agentic AI. SAP is also fully committed to this. There are plenty of agents on their way to automate finance processes. Major leaps forward are expected this year. Does this make SAP late compared to other software suppliers? Yes and no. The type of autonomous agent that SAP has in mind is fully integrated into finance processes, while what is already on the market deals with less sensitive processes and data. In the area where SAP is active, it is therefore simply ahead of the curve according to the roadmap.
The use cases for AI in ERP
If we look at what artificial intelligence can do in ERP processes, we automatically arrive at generative AI. SAP has been working with machine learning and AI tools for years, but the advent of generative technologies has changed the perspective. One of the most common applications is Enterprise Service Management, where AI analyzes incoming documents. Information can be extracted from invoices, complaints, or unpaid bills, and the technology can conclude to trigger an action. The technology reads these documents, links them to the correct transactions, analyzes the content, and suggests actions. This application is particularly popular with service and help desks and HR and sales departments.
AI is also increasingly being used in the logistics sector. There, it can read and process shipping documents, which can sometimes run to hundreds of pages. AI delivers significant efficiency gains, as it can plough through the text in seconds or minutes. Based on these documents, actions are proposed that can improve logistics processes.
Why certain professional groups are leading the way
SAP also acknowledges that the first applications of AI in ERP systems were mainly visible in HR and sales. Now, expansion to financial processes needs to gain momentum. According to SAP, there are two reasons for this: these processes are easier to standardize and represent a large part of the daily work within organizations.
“HR processes are already fairly standardized by definition,” says CPO Industries, Globalization Engineering & CMO SAP Business Suite Van Rossum. “These were also the first applications to move to the cloud.” This makes these processes very suitable for AI applications. Standardization ensures a consistent data model essential for effective AI implementation.
Another factor is that relatively few people in a company are involved in finance. They are usually experts who perform specific tasks. AI tools can increase their productivity, while other employees who are not directly involved in financial processes can also benefit from improved workflows.
The cloud as a prerequisite for AI
One aspect we touched on during our conversation with Van Rossum is that SAP makes the new AI functionalities available almost exclusively via cloud solutions. This means that older ECC systems are left out in the cold. There are still plenty of users, and there are no hard figures on the ratio of ECC to S/4HANA, but in this case, they are lagging behind.
From a technical point of view, the decision to bring AI innovation to newer versions is also understandable. On-premises ECC users often have heavily customized ERP systems, which makes it difficult to implement and use AI effectively. Opting for a clean-core or fit-to-standard ERP (S/4 public or private cloud) does make it possible to receive regular updates, which allows you to implement and consume the latest AI.
For SAP, this is also a conscious strategic choice. Van Rossum explains that SAP currently has several core products. “You have the private cloud ERP solution based on the S/4 product lines and a public cloud version with a hyperscaler. Depending on a company’s needs and preferences, one of the two will be the most suitable. And with the true SaaS solution, the customer gets the added benefit of access to the latest innovations via every six-month upgrade.”
The decision to deliver AI only on cloud solutions has other reasons besides a possible extra incentive to move towards S/4HANA. Cloud solutions’ standardization makes it easier to implement AI tools effectively. In addition, it is easier for SAP to implement innovations in a cloud environment because updates can be rolled out centrally.
From a few AI agents to collaborative agents
With S/4HANA, you automatically have some AI options for automating processes. Meanwhile, work on the future of agentic AI is in full swing. One of the most important things to watch at SAP is the multi-agent system concept, in which multiple AI agents work together to perform more complex tasks. In the current phase, single agents are mainly used to optimize one specific task, such as collecting payments.
SAP envisions a future in which, similar to how people work together, AI agents will also communicate with each other. “Right now, it’s basically agents talking to people,” Van Rossum observes. The expectation is that this will evolve into a model in which agents divide tasks among themselves and pass them on to each other. However, human involvement will remain essential, according to SAP.
This means a shift in how tasks are performed, but not necessarily the disappearance of jobs. “Since the industrial revolution, there has never been an innovation or shift that has rendered people useless,” says Van Rossum. “But of course, if you look back 100 years, there were professions that no longer exist today, and that will certainly happen in this transition as well.”
Optimizing ABAP code
Van Rossum has held a senior position in the Cloud ERP team for quite some time. He has seen a lot of changes in how SAP systems are built, implemented, and optimized. This is reflected in the optimization of ABAP code, the programming language used within SAP. SAP has placed all its ABAP code in a large language model (LLM) for analysis and optimization.
“When I started at SAP 25 years ago, I wrote my own ABAP code and you just started from scratch,” Van Rossum explains. “But now there is so much code, all written at some point. The people who wrote it are no longer there.” With the help of AI, existing code blocks can be optimized and new code can be generated, securing the future of ABAP development. It remains a useful language for optimally configuring the ERP system, and that knowledge is needed, but the need is too great not to link AI to it for efficiency gains.
Migration to new systems accelerates
From a business perspective, the arrival of AI functionalities for SAP is beneficial. It encourages companies to switch from older ERP systems to modern cloud solutions more quickly. Although Van Rossum cannot share specific figures when asked, he does make it clear that an increasing number of customers are migrating directly to the public cloud ERP solution instead of first moving to a private cloud environment.
“We are seeing more and more of these customers actually jumping straight to the public cloud ERP solution, precisely to gain the advantage of ‘fit for standard’. Companies can easily consume these cloud innovations. AI also plays a very important role in this, because it makes things easier,” explains Van Rossum.
The biggest challenge in these migrations is not so much the technology, but the organizational changes that come with it. The speed at which organizations can make the switch is mainly determined by their ability to adapt business processes to the standards offered by the cloud solution, as is the philosophy of S/4HANA.
All in all, it is clear to SAP users that AI can play an increasingly important role in their daily work. Whether it’s analyzing documents, optimizing code, or automating routine tasks, the integration of AI into ERP processes promises significant efficiency gains.
Tip: SAP brings all data together in one place with Business Data Cloud